I ended up reprinting the sensor bed twice. It turns out 3D printing the entire thing with supports, while it will print, it's so hard to get the supports out. So I had to rework the design and split it into three parts.

I also interfaced with all the parts so I can pull data from them.

The servo gear unfortunately tilts it to the right slightly due to the gear raising the whole body up slightly.

So yeah, I'll clean up the wires and then I'll do the programming. I still have to design the outer housing and bottom mount. Then bridge it with the tail dragger robot that has its own separate electronics. I'll get it done (physical part) over the weekend for sure.

I did make a mistake, that peg that holds the LED blocked the ribbon cable for the camera. So the camera is mounted upside down... the lens is supposed to be centered with the other sensors. There's an IMU that's not visible but it's in the middle of the rotating plane directly behind the camera.

The other thing that's funny is I'm using this TFmini-s library by budryerson and it's so accurate. His library is meant for the TFmini plus but it works with the S too. In the past the TFmini-s was only good for 30cm or farther... now it can scan as close as 1 cm... so the VL53L0X is kind of redundant but it has a wide FOV (25 deg. vs 3.6). So it's still useful.

I'll use the physical measurements to double check the OpenCV blog scanning and then through a web API I'll check if what the robot saw vs. what decision it made was a good or bad and somehow use that to make it better.